Mass transfer between fluids as a mechanism for seismic wave attenuation: experimental evidence from water–CO2 saturated sandstones

Abstract

Summary Seismic waves are typically assumed to propagate fast enough through a porous rock saturated with multiple fluid phases such that the interaction between the fluids can be considered adiabatic, or thermodynamically unrelaxed. However, at low gas saturations and when the gas is present in the form of microscopic bubbles the fluid mixture may in fact be thermodynamically relaxed at seismic frequencies. The effective fluid is then significantly more compressible. A transition from a thermodynamically relaxed to unrelaxed state of the fluids will be accompanied by frequency dependent attenuation of the wave in response to heat and/or mass transfer. We conducted forced oscillation experiments on two partially saturated sandstone samples to measure frequency dependent attenuation and modulus dispersion at seismic frequencies (< 1000 Hz). For CO2 saturations of 0.1-0.2% we observe significant attenuation and dispersion in the bulk and shear modulus, with an attenuation peak at ∼100 Hz. The bulk modulus was significantly lower than the prediction by Gassmann-Wood fluid substitution, which assumes that the fluids are thermodynamically unrelaxed. Numerical simulations in poroelastic media further indicate that a partially drained boundary condition does not adequately explain the observed attenuation and dispersion, particularly in the shear modulus. Numerical simulations at the microscopic scale support the notion that pore-scale heterogeneities could explain the observed shear attenuation and dispersion, since an external shear deformation can cause local compressions of the pore space. The observed attenuation and dispersion are interpreted to be predominantly due to a transition from a thermodynamically relaxed to unrelaxed state of the saturating fluids.

Influence of Fluid Distribution on Seismic Dispersion and Attenuation in Partially Saturated Limestone

“…(1995), Akin and Kovscek (2003), and Chapman et al. (2021, 2022). The preprocessing steps are the followings: (a) we first normalized the image of the fully saturated sample to the dry sample, using the “Normalize Grayscale” model in Avizo (Lin et al., 2017); (b) in a second step, we made a difference between the gray values in the images of the fully water‐saturated sample and the dry sample, that is, $$C{T}_{Water}-C{T}_{Air}$$; and (c) the porosity distribution is obtained as $$$\phi =\frac{C{T}_{Water}-C{T}_{Air}}{{\Theta}}$$$ where $${\Theta}\,$$ refer to the normalization coefficient.…”

“…The samples were wrapped in plastic foil to avoid drying during the scans. The CT scanning results can be used to estimate the porosity and fluid distribution, following the methodology given in Cadoret et al (1995), Akin and Kovscek (2003), and Chapman et al (2021Chapman et al ( , 2022. The preprocessing steps are the followings: (a) we first normalized the image of the fully saturated sample to the dry sample, using the "Normalize Grayscale" model in Avizo (Lin et al, 2017); (b) in a second step, we made a difference between the gray values in the images of the fully water-saturated sample and the dry sample, that is, 𝐴𝐴 𝐴𝐴𝐴𝐴𝑊𝑊 𝑊𝑊𝑊𝑊𝑊𝑊𝑊𝑊 − 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝑊𝑊 ; and (c) the porosity distribution is obtained as…”

On Frequency-Dependent Rock Experiments: A Comparative Review

“…Numerical studies supplementing experimental ones are becoming customary. Experimentally proven by Tisato and Madonna [102], Quintal et al [191] and Quintal et al [192] were the numerical impetus of Chapman and Quintal [85], Tisato and Quintal [106], Chapman et al [127,134], Gallagher et al [149], Tisato and Quintal [166], Kuteynikova et al [167], Quintal et al [193], Hunziker et al [194], Quintal et al [195], Alkhimenkov et al [196], Lissa et al [197][198][199][200], and Alkhimenkov and Quintal [201,202] in terms of attenuation modelling. Zhang and Toksöz [203], Das et al [204], and Jänicke et al [205] used similar Digital Rock Physics (DRP) approaches to study attenuation mechanisms and fluid-solid interactions.…”

“…Lu [133] measured axial and shear strains of artificial specimens (3D printed and otherwise) at different stress and loading conditions from 0.01 to 20 Hz using a pair of Laser Displacement Sensors (LDS). Chapman et al [134] questioned the assumed adiabatic (thermodynamically unrelaxed) and instead argued for isothermal (thermodynamically relaxed) interaction between multiple fluid phases by studying microscopic gas bubbles. Evident E, K, and G dispersion and…”